I cast my lot here with Valour and other similar thinkers. The fact is that fingering can make or break a performance in regards to fluidity. Within each phrase, fingering can enhance or destroy the musical line and objective. Also, the quicker you achieve a truly well-thought out, workable fingering, the faster you can also unlearn a less-than- optimal fingering of convenience.
In learning a new piece, there are obvious clues when fingering goes awry. Firstly, if you run out of fingers in passage work, you need to stop, observe, consider, try, retry, and figure out a viable solution. Secondly, undesirable, sudden, and illlogical shifts in hand position can also indicate a serious problem with fingering. Unnecessary shifts, bumps, and jumps will never add to the aesthetics of the sound. Thirdly, when conventional fingerings are simply not achieving the aim, it's time to look at less ordinary solutions--like Chopin's crossing the 4th finger over the 5th, for example. Fourthly, a particular fingering may be inconsistent with touch. For instance, if a finger legato is just not working properly, the fingering must be scrutinized then and there before bad habits set in. Often you can choose a fingering more easily by working the passage backwards--try it! There are other indicators of the need to consider fingerings, but these are some common ones.
Often, deciding on a proper fingering is all about preparing the hand in advance to be in the right place at the right time. For example, it might be important to be at the top of a phrase on a particular finger to facilitate the receding portion of that phrase. Or, you might need to use a clever fingering to prepare the hand for a chord. A good fingering can greatly assist in preparing the hand for a wide leap. And it can help in smoothing out a difficult arpeggio figuration. Players who don't prepare in this fashion must cope with nasty, but avoidable, surprises instead.
In learning a new piece, the decisions I reach in the matters of fingering are not necessarily final and binding. Rather I keep them as tentative choices. As I become more familiar with the requirements of choreographing the hands, further alterations in fingering might make perfect sense. For example, there might be a great justification for distributing music between the hands in a measure, thereby totally changing the original fingering selected.
Finally, I always use urtext or the best editions I can find. Even there, I do not take written fingerings as gospel (except when indicated by the composer himself), even from a famous pedagogue and editor like Joseffy. Because everyone's hand is structured slightly differently, what worked perfectly for Joseffy might not adapt well to my own hand or someone else's--meaning close analyis is in order. So there are many clues and considerations. I strongly feel that those who want to simply see how the notes fall under the hand and pell mell allow a fingering of convenience to occur inconsistent with performance practice are making a serious mistake in honing their ability to play that piece to their fullest potential and in accord with the composer's expectations. Sometimes the lazy way out will work; but too often it doesn't.